M2cai16-tool-locations (2024)

boxes = target['boxes'].int() labels = target['labels'] class_names = dataset.CLASSES

# 16 tool classes (example; adjust to your annotation file) CLASSES = [ 'background', 'grasper', 'scissors', 'hook', 'clipper', 'irrigator', 'specimen_bag', 'bipolar', 'hook_electrode', 'trocars', 'stapler', 'suction', 'clip_applier', 'vessel_sealer', 'ligasure', 'ultrasonic', 'other' ]

import json import os from PIL import Image import torch from torch.utils.data import Dataset from torchvision.ops import box_convert class M2CAI16ToolLocations(Dataset): """Dataset for m2cai16-tool-locations bounding box annotations."""

def __init__(self, root_dir, transform=None): self.root_dir = root_dir self.transform = transform self.samples = [] # Collect all (frame_path, annotation_path) pairs ann_dir = os.path.join(root_dir, 'annotations') for ann_file in os.listdir(ann_dir): if not ann_file.endswith('.json'): continue ann_path = os.path.join(ann_dir, ann_file) video_id = ann_file.replace('.json', '') frame_dir = os.path.join(root_dir, 'frames', video_id) with open(ann_path, 'r') as f: annotations = json.load(f) for frame_name, boxes_info in annotations.items(): frame_path = os.path.join(frame_dir, frame_name) if os.path.exists(frame_path): self.samples.append((frame_path, boxes_info)) m2cai16-tool-locations

def __len__(self): return len(self.samples)

# Draw boxes img_with_boxes = draw_bounding_boxes(img, boxes, labels=[class_names[l] for l in labels], colors='red', width=2) plt.figure(figsize=(10, 8)) plt.imshow(img_with_boxes.permute(1,2,0)) plt.axis('off') plt.title(f"Frame {idx} — {len(boxes)} tools detected") plt.show() dataset = M2CAI16ToolLocations('./m2cai16-tool-locations') show_annotations(dataset, idx=0) 4. Useful Preprocessing for Training Convert to COCO format (for Detectron2, MMDetection, etc.):

This dataset is designed for (bounding boxes) in laparoscopic cholecystectomy videos. It contains annotations for 16 tools, including their positions in video frames. 1. Dataset Overview & Utility Purpose : Train object detection models (e.g., YOLO, Faster R-CNN, DETR) to locate surgical instruments in real-time. boxes = target['boxes']

import matplotlib.pyplot as plt from torchvision.utils import draw_bounding_boxes from torchvision.transforms import ToTensor def show_annotations(dataset, idx=0): img, target = dataset[idx] if isinstance(img, torch.Tensor): img = (img * 255).byte() if img.max() <= 1 else img else: img = ToTensor()(img).byte()

yolo detect train data=m2cai16.yaml model=yolov8n.pt epochs=100 imgsz=640 Example m2cai16.yaml :

path: ./m2cai16-tool-locations train: images/train val: images/val nc: 16 names: ['grasper','scissors','hook','clipper','irrigator','specimen_bag','bipolar','hook_electrode','trocars','stapler','suction','clip_applier','vessel_sealer','ligasure','ultrasonic','other'] This guide gives you a production‑ready starting point for loading, visualizing, converting, and training on the dataset. Adjust class names and annotation JSON structure based on your exact dataset version. Adjust class names and annotation JSON structure based

m2cai16-tool-locations/ annotations/ video01.json # or .xml / .txt video02.json frames/ video01/ frame_000001.jpg ... Here’s a robust parser using and torchvision :

def __getitem__(self, idx): img_path, ann = self.samples[idx] image = Image.open(img_path).convert('RGB') # Parse annotations: list of [x1, y1, x2, y2, class_id] boxes = [] labels = [] for obj in ann.get('objects', []): x1, y1, x2, y2 = obj['bbox'] # absolute pixel coords label = self.CLASSES.index(obj['class_name']) boxes.append([x1, y1, x2, y2]) labels.append(label) boxes = torch.as_tensor(boxes, dtype=torch.float32) labels = torch.as_tensor(labels, dtype=torch.int64) image_id = torch.tensor([idx]) area = (boxes[:, 3] - boxes[:, 1]) * (boxes[:, 2] - boxes[:, 0]) iscrowd = torch.zeros((len(boxes),), dtype=torch.int64) target = { 'boxes': boxes, 'labels': labels, 'image_id': image_id, 'area': area, 'iscrowd': iscrowd } if self.transform: image, target = self.transform(image, target) return image, target Use matplotlib and torchvision.utils.draw_bounding_boxes :